if (!requireNamespace("groundhog", quietly = TRUE)) {
install.packages("groundhog")
library("groundhog")
}
pkgs <- c("magrittr", "data.table", "stringr", "Rmisc", "ggplot2", "patchwork",
"rstatix", "ggpubr", "glue", "lubridate", "plyr",
"lmtest", "sandwich", "stargazer" # packages for regression tables
)
groundhog::groundhog.library(pkg = pkgs,
date = "2023-09-25")6 Online Appendix C
6.1 Setup
6.1.1 Install Packages
We install the following packages using the groundhog package manager to increase computational reproducibility.
6.1.2 Read Data
# data <- data.table::fread(file = "../data/processed/full.csv")
data <- readRDS(file="../data/processed/full.Rda")6.1.3 Manipulate Data
[Shall we add this code chunk to the pre-processing?]
data[, communication := as.factor(communication)]
data[, communication := factor(communication, levels = c("point", "both","interval"))]
data[, stage := as.factor(stage)]
data[, stage := factor(stage, levels = c("1", "2"))]cols <- str_subset(string = names(data), pattern = "E\\d+")
data[,
str_replace_all(string = cols, pattern = "^E", replacement = "m") := lapply(.SD, function(x) x / 100),
.SDcols = cols]
data[, p1 := 0.5 + (m1 - m23)/ (6*(mc-ms))]
data[, p2 := 0.5 + (m2 - m13)/ (6*(mc-ms))]
data[, p3 := 0.5 + (m3 - m12)/ (6*(mc-ms))]wide_data <- data.table::dcast(data,
participant.label + surprise + communication ~ stage,
value.var = c("E1", "E2", "E3", "E12", "E23","E13", "b", "a"))
setorder(wide_data, surprise, communication, participant.label)6.1.4 Design
We define some design features in the following:
colors <- c("#F3B05C", "#1E4A75", "#65B5C0", "#AD5E21")
layout <- theme(panel.background = element_rect(fill = "white"),
legend.key = element_rect(fill = "white"),
panel.grid.major.y = element_line(colour = "grey",
linewidth = 0.25),
axis.ticks.y = element_blank(),
panel.grid.major.x = element_blank(),
axis.line.x.bottom = element_line(colour = "#000000",
linewidth = 0.5),
axis.line.y.left = element_blank(),
plot.title = element_text(size = rel(1))
)6.1.5 Helper Function
To create Figure 6.1 (and the other figures), we create a wrapper function that we can call several times. Figures such as Figure 6.1 consist of eight panels, that are relatively similar. We thus, save both space and sources of error by creating a wrapper function plot_scatter() that creates scatter plots. To annotate and scale these plots, we also create a wrapper function for plot_scatter() called plot_scatter_with_defaults(). The latter function is called in the script below and contains some hard-coded defaults that determine annotations and the scales, i.e., where the annotations are placed.
plot_scatter <- function(outcome.y = "b_2",
outcome.x = "b_1",
DT = wide_data[surprise == FALSE & communication == "both"],
breaks = seq(from = -1, to = 1, by = 0.5),
limits = c(-1, 1),
# manual annotation positions:
x1 = 0.9,
x2 = 0.3,
x3 = -0.6,
x4 = 0.6,
y1 = - 0.115,
y2 = 1.05,
y3 = 0.7,
y4 = -0.7){
# Percentage Annotation
p1 <- scales::percent(x = sum(DT[,..outcome.y] > DT[,..outcome.x], na.rm = TRUE) / DT[, .N],
accuracy = 1)
p2 <- scales::percent(x = sum(DT[,..outcome.y] < DT[,..outcome.x], na.rm = TRUE) / DT[, .N],
accuracy = 1)
# Spearman Correlation for Subtitle
ro.value <- round(x = cor(DT[,..outcome.x],
DT[,..outcome.y],
method = c("spearman")),
digits = 2)
# Plot
plot <- ggplot(data = DT,
mapping = aes(x = get(outcome.x),
y = get(outcome.y))) +
geom_count(shape = 1, show.legend = FALSE) +
# add axes and diagonal
geom_vline(xintercept = 0) +
geom_hline(yintercept = 0) +
geom_abline(intercept = 0, slope = 1, linewidth = 0.6) +
# remove all the other lines
theme(line = element_blank(),
panel.background = element_rect(fill = "white",
colour = "white",
linetype = "solid")) +
# define axis labels
labs(x = as.expression(bquote(rho~"="~.(ro.value))),
y = "") +
# add annotations
geom_text(x = x1, y = y1, label = "Part 1") +
geom_text(x = x2, y = y2, label = "Part 2") +
geom_text(x = x3, y = y3, label = p1) +
geom_text(x = x4, y = y4, label = p2) +
# define consistent axis breaks etc.
scale_x_continuous(breaks = breaks, lim = limits) +
scale_y_continuous(breaks = breaks, lim = limits)
plot
}
# Wrapper function with a consistent parameter handling approach
plot_scatter_with_defaults <- function(DT, figure = "oc1") {
# Define defaults for each set
defaults <- list(
oc1 = list(outcome.y = "b_2", outcome.x = "b_1",
breaks = seq(from = -1, to = 1, by = 0.5),
limits = c(-1, 1),
x1 = 0.9, x2 = 0.3, x3 = -0.6, x4 = 0.6,
y1 = -0.115, y2 = 1.05, y3 = 0.7, y4 = -0.7),
oc2 = list(outcome.y = "a_2", outcome.x = "a_1",
breaks = seq(from = -2, to = 4, by = 1),
limits = c(-2, 4),
x1 = 3.4, x2 = 0.8, x3 = -1.5, x4 = 1.5,
y1 = -0.4, y2 = 0.39, y3 = 1.9, y4 = -1.9),
oc3 = list(outcome.y = "E1_2", outcome.x = "E1_1",
breaks = seq(from = 0, to = 100, by = 50),
limits = c(0, 101),
x1 = 90, x2 = 15, x3 = 15, x4 = 90,
y1 = 5, y2 = 100, y3 = 80, y4 = 20),
oc4 = list(outcome.y = "E2_2", outcome.x = "E2_1",
breaks = seq(from = 0, to = 100, by = 50),
limits = c(0, 101),
x1 = 90, x2 = 15, x3 = 15, x4 = 90,
y1 = 5, y2 = 100, y3 = 80, y4 = 20),
oc5 = list(outcome.y = "E3_2", outcome.x = "E3_1",
breaks = seq(from = 0, to = 100, by = 50),
limits = c(0, 101),
x1 = 90, x2 = 15, x3 = 15, x4 = 90,
y1 = 5, y2 = 100, y3 = 80, y4 = 20),
oc6 = list(outcome.y = "E12_2", outcome.x = "E12_1",
breaks = seq(from = 0, to = 100, by = 50),
limits = c(0, 101),
x1 = 90, x2 = 15, x3 = 15, x4 = 90,
y1 = 5, y2 = 100, y3 = 80, y4 = 20),
oc7 = list(outcome.y = "E13_2", outcome.x = "E13_1",
breaks = seq(from = 0, to = 100, by = 50),
limits = c(0, 101),
x1 = 90, x2 = 15, x3 = 15, x4 = 90,
y1 = 5, y2 = 100, y3 = 80, y4 = 20),
oc8 = list(outcome.y = "E23_2", outcome.x = "E23_1",
breaks = seq(from = 0, to = 100, by = 50),
limits = c(0, 101),
x1 = 90, x2 = 15, x3 = 15, x4 = 90,
y1 = 5, y2 = 100, y3 = 80, y4 = 20)
)
# Remove the non-required default params before calling the plot function
default_params <- defaults[[figure]]
# Combine all parameters
args <- c(list(DT = DT), default_params)
# Call the original plot function with the combined arguments
do.call(plot_scatter, args)
}6.2 Figure OC.1
In what follows, we will call the custom functions defined above.
top1 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE], figure = "oc1")
top2 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE], figure = "oc1")
mid1 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE & communication == "point"], figure = "oc1")
mid2 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE & communication == "interval"], figure = "oc1")
mid3 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE & communication == "both"], figure = "oc1")
bot1 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE & communication == "point"], figure = "oc1")
bot2 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE & communication == "interval"], figure = "oc1")
bot3 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE & communication == "both"], figure = "oc1")
((top1|top2) / (mid1 | mid2 | mid3) / (bot1 | bot2 | bot3) & theme(legend.position = "bottom")) + plot_layout(guides = "collect")
6.3 Figure OC.2
top1 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE], figure = "oc2")
top2 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE], figure = "oc2")
mid1 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE & communication == "point"], figure = "oc2")
mid2 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE & communication == "interval"], figure = "oc2")
mid3 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE & communication == "both"], figure = "oc2")
bot1 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE & communication == "point"], figure = "oc2")
bot2 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE & communication == "interval"], figure = "oc2")
bot3 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE & communication == "both"], figure = "oc2")
((top1|top2) / (mid1 | mid2 | mid3) / (bot1 | bot2 | bot3) & theme(legend.position = "bottom")) + plot_layout(guides = "collect")
6.4 Figure OC.3
top1 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE], figure = "oc3")
top2 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE], figure = "oc3")
mid1 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE & communication == "point"], figure = "oc3")
mid2 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE & communication == "interval"], figure = "oc3")
mid3 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE & communication == "both"], figure = "oc3")
bot1 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE & communication == "point"], figure = "oc3")
bot2 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE & communication == "interval"], figure = "oc3")
bot3 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE & communication == "both"], figure = "oc3")
((top1|top2) / (mid1 | mid2 | mid3) / (bot1 | bot2 | bot3) & theme(legend.position = "bottom")) + plot_layout(guides = "collect")
6.5 Figure OC.4
top1 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE], figure = "oc4")
top2 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE], figure = "oc4")
mid1 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE & communication == "point"], figure = "oc4")
mid2 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE & communication == "interval"], figure = "oc4")
mid3 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE & communication == "both"], figure = "oc4")
bot1 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE & communication == "point"], figure = "oc4")
bot2 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE & communication == "interval"], figure = "oc4")
bot3 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE & communication == "both"], figure = "oc4")
((top1|top2) / (mid1 | mid2 | mid3) / (bot1 | bot2 | bot3) & theme(legend.position = "bottom")) + plot_layout(guides = "collect")
6.6 Figure OC.5
top1 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE], figure = "oc5")
top2 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE], figure = "oc5")
mid1 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE & communication == "point"], figure = "oc5")
mid2 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE & communication == "interval"], figure = "oc5")
mid3 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE & communication == "both"], figure = "oc5")
bot1 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE & communication == "point"], figure = "oc5")
bot2 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE & communication == "interval"], figure = "oc5")
bot3 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE & communication == "both"], figure = "oc5")
((top1|top2) / (mid1 | mid2 | mid3) / (bot1 | bot2 | bot3) & theme(legend.position = "bottom")) + plot_layout(guides = "collect")
6.7 Figure OC.6
top1 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE], figure = "oc6")
top2 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE], figure = "oc6")
mid1 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE & communication == "point"], figure = "oc6")
mid2 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE & communication == "interval"], figure = "oc6")
mid3 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE & communication == "both"], figure = "oc6")
bot1 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE & communication == "point"], figure = "oc6")
bot2 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE & communication == "interval"], figure = "oc6")
bot3 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE & communication == "both"], figure = "oc6")
((top1|top2) / (mid1 | mid2 | mid3) / (bot1 | bot2 | bot3) & theme(legend.position = "bottom")) + plot_layout(guides = "collect")
6.8 Figure OC.7
top1 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE], figure = "oc7")
top2 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE], figure = "oc7")
mid1 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE & communication == "point"], figure = "oc7")
mid2 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE & communication == "interval"], figure = "oc7")
mid3 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE & communication == "both"], figure = "oc7")
bot1 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE & communication == "point"], figure = "oc7")
bot2 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE & communication == "interval"], figure = "oc7")
bot3 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE & communication == "both"], figure = "oc7")
((top1|top2) / (mid1 | mid2 | mid3) / (bot1 | bot2 | bot3) & theme(legend.position = "bottom")) + plot_layout(guides = "collect")
6.9 Figure OC.8
top1 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE], figure = "oc8")
top2 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE], figure = "oc8")
mid1 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE & communication == "point"], figure = "oc8")
mid2 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE & communication == "interval"], figure = "oc8")
mid3 <- plot_scatter_with_defaults(DT = wide_data[surprise == FALSE & communication == "both"], figure = "oc8")
bot1 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE & communication == "point"], figure = "oc8")
bot2 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE & communication == "interval"], figure = "oc8")
bot3 <- plot_scatter_with_defaults(DT = wide_data[surprise == TRUE & communication == "both"], figure = "oc8")
((top1|top2) / (mid1 | mid2 | mid3) / (bot1 | bot2 | bot3) & theme(legend.position = "bottom")) + plot_layout(guides = "collect")